Michael Gazzaniga’s Neuro-cognitive Antireductionism
and the Challenge of Neo-mechanistic Reduction
Diego Azevedo Leite
(α)Ricevuto: 1 febbraio 2018; accettato: 25 giugno 2018
█ Abstract Michael Gazzaniga, a prominent cognitive neuroscientist, has argued against reductionist
ac-counts of cognition. Instead, Gazzaniga defends a form of non-reductive physicalism: epistemological neuro-cognitive non-reductionism and ontological monist physicalism. His position is motivated by the theses that: (1) cognitive phenomena can be realized by multiple neural systems; (2) many outcomes of these systems are unpredictable; and (3) multi-level explanations are required. Epistemological neuro-cognitive non-reductionism is presented as the most appropriate stance to account for the way in which phenomena should be explained in cognitive neuroscience. In this paper, I argue, however, that a recent form of (ontological and epistemological) neuro-cognitive reductionism, namely neo-mechanistic reduc-tionism accounts for the arguments presented by Gazzaniga. Thus, the theory offers a more consistent and well-articulated view of the relationship between cognitive and neural phenomena that is specifically compatible with the explanatory strategies and aims of contemporary cognitive neuroscience.
KEYWORDS: Neo-mechanistic Philosophy; Michael Gazzaniga; Non-reductionism; Reductionism; Philos-ophy of Cognitive Neuroscience
█ Riassunto L’antiriduzionismo neurocognitivo di Michael Gazzaniga e la sfida della riduzione
neomeccani-cista – Uno dei più importanti neuroscienziati dei nostri tempi, Michael Gazzaniga, si è schierato contro una concezione riduzionista della cognizione. Al contrario Gazzaniga difende una forma di fisicalismo non-riduzionistico che risulta dalla combinazione, sul piano epistemologico, di un anti-riduzionismo co-gnitivo e, sul piano ontologico, di un monismo fisicalista. La sua posizione è motivata dalla tesi per cui (1) i fenomeni cognitivi possono essere realizzati da molteplici sistemi neurali; (2) molti esiti di tali sistemi non si possono prevedere; (3) e sono pertanto necessarie spiegazioni a livelli plurimi. La concezione pre-sentata come più adeguata per dare conto sul piano epistemologico di come i fenomeni dovrebbero essere spiegati all’interno delle neuroscienze cognitive è una forma di anti-riduzionismo neuro-cognitivo. In questo articolo si sostiene tuttavia che una recente forma di riduzionismo neuro-cognitivo (epistemologi-co ed ontologi(epistemologi-co) può dare (epistemologi-conto degli argomenti presentati da Gazzaniga. Si tratta di una teoria che offre una interpretazione maggiormente coerente ed articolata della relazione fra fenomeni cognitivi e neurali e che offre un modello di spiegazione compatibile con gli scopi esplicativi delle neuroscienze cognitive con-temporanee.
PAROLE CHIAVE: Filosofia neomeccanicista; Michael Gazzaniga; Nonriduzionismo; Riduzionismo; Filoso-fia della neuroscienza cognitiva
RICERCHE
(α)Dipartimento di Psicologia e Scienze Cognitive, Università degli Studi di Trento, Corso Bettini, 84 - 38068 Rovereto (I)
E-mail: diego.azevedo.leite@gmail.com () Creative Commons - Attribuzione - 4.0 Internazionale
MICHAEL GAZZANIGA IS ONE OF the most distinguished researchers in the contempo-rary field of cognitive neuroscience.1 Some authors even consider him to be the “father” of this research area.2 Thanks to his numer-ous theoretical and empirical contributions to the field over a span of more than fifty years, he has had an enormous influence on many significant debates, e.g. the effects of split-brain surgery on cognitive phenomena such as visual consciousness and language;3 the relationships between the human brain and moral beliefs, and between neuroscience and ethics;4 the relation between neurosci-ence, free will and law;5 the possible evolu-tion of certain brain structures;6 and how the brain generates social beliefs.7 His theories are representative of many fundamental ide-as in these areide-as of research.
Gazzaniga’s important impact on the field is also due to the fact that he is one of the few contemporary cognitive neuroscientists who tries to explicitly and directly address the de-bate on the so-called mind-body problem (or neuro-cognitive relation),8 a problem which tacitly underlies all discussion in cognitive neuroscience, since at the most fundamental theoretical level this discipline interrogates the relationship between neural systems and cognition.9 Indeed, this line of research is de-scribed as a «combined study of mind and brain»10 whose ultimate aim is «to provide a brain-based account of cognition».11 A dis-cussion of Gazzaniga’s perspective is not only important for understanding his particular position on this issue, but also for more broadly examining the conception of the re-lationship between brain and cognition im-plicit in the very structure of the neuro-cognitive research informed by his views.
In order to understand, characterize and explain cognition, a number of authors con-sider reductive approaches attractive. From an ontological point of view, these approaches usually imply a simpler and more integrated (unitary) conception of the entities and the processes that science and philosophy investi-gate.12 From an epistemological point of view,
reductive approaches usually also help to ad-vance scientific integration, avoiding gaps be-tween scientific domains and theories.13
However, other authors consider reduc-tive approaches problematic especially be-cause of their implications for the way in which cognitive phenomena are character-ized and for preserving the autonomy of psy-chological theories in the face of neuroscien-tific theories.14 In fact, it is extremely difficult to pursue the reductionist ambition to inte-grate or even to link traditional psychological concepts and explanations with neurophysio-logical concepts and explanations.
Gazzaniga belongs to a group of authors who expressly embrace an epistemological neuro-cognitive non-reductionist position. He suggests a version of a neuro-cognitive non-reductive explanation. This solution is purported to overcome problems with the neuro-cognitive epistemological reductionist view and to offer, at the same time, a plausi-ble ontological and epistemological account of the relationship between cognition and the brain that is appropriate for the ultimate ex-planatory aims of cognitive neuroscience. In this paper, I will examine Gazzaniga’s view and show that even though he attempts to construct an epistemological non-reductive account of cognition, his view is completely compatible with a particular kind of contem-porary ontological and epistemological re-ductionist position.
I will discuss Gazzaniga’s position on re-ducibility, as well as the reasons why he con-siders epistemological non-reductionism to be the most suitable approach to explana-tions of (human) neuro-cognitive phenome-na. The author understands epistemological neuro-cognitive reductionism as intertheo-retical relations of derivation when identity relations between terms are constructed. He claims that epistemological reduction so con-sidered is problematic for three reasons: (1) cognitive phenomena are realized by multiple systems; (2) the outcomes of neuro-cognitive systems are unpredictable; and (3) multi-level explanations are indispensable. On this
basis, I will show that Gazzaniga’s defense of the epistemological non-reductionist view is problematic and is actually compatible with a particular kind of contemporary ontological and epistemological neuro-cognitive reduc-tionism. Gazzaniga believes that epistemo-logical neuro-cognitive non-reductionism is the only position that can account for all three of these aspects. However, I argue that multiple realizability, unpredictability, and multi-level explanations, as framed by Gaz-zaniga, can be accommodated by the neo-mechanistic neuro-cognitive reductionist ap-proach, which is also compatible with the ex-planatory strategies and ultimate aims of contemporary cognitive neuroscience. █ Gazzaniga’s defense of non-reductionism
According to Gazzaniga, evidence from neurologically impaired patients, especially split-brain patients, and recent neuroimaging data show that the human brain is organized into local circuits, specialized for specific functions, known as modules.15 These cir-cuits run simultaneously in parallel, are dis-tributed all over the brain and process differ-ent inputs through various kinds of automat-ic computation. These different systems con-tain millions of networks that are connected to and affect each other. It is these different systems that implement cognitive processes.16 Cognitive processes are thus understood as information processes implemented by neural systems; more specifically, as information processes realized by automatic computations.
The same applies to those mental phe-nomena we can consciously access, e.g. per-ceptions, beliefs, desires, intentions etc., which are “enabled” individually by a num-ber of local neural systems and subsystems.17 Although consciousness appears to be uni-fied, it is generated by these vastly separate systems. Somehow these systems are able to integrate and unify the information they pro-cess and get it to the level of consciousness. This happens when the brain “considers” the information they provide to be most
im-portant at a specific moment. Thus, Gazza-niga’s view essentially seems to be that what-ever experience a person happens to be con-scious of at a particular moment, this is the one that became dominant and got «the prize of conscious recognition».18 The gene-sis of beliefs is also explained as the result of a brain process. More specifically, beliefs are said to be produced by a special module in the left hemisphere, called the interpreter. The function of this mechanism consists in processing all the inputs of other systems in the brain in order to “interpret” them, i.e. to create a personal, consistent narrative of our past (the prior actions of our nervous system). This system looks for patterns: it is driven to «seek explanations or causes for events» and «hypothesize about the structure of the world».19 It is the “interpretation” of the in-formation made available by this brain mod-ule that actually creates human beliefs: «[...] there is a system in the left hemisphere of the brain [...], which produces beliefs [...]».20
The view that thinking is a form of in-formation processing based on computations (on symbolic or sub-symbolic representa-tions) provides an idea of how a purely phys-ical system might bring about specific “cogni-tive processes” (e.g. playing chess, recogniz-ing faces, and reproducrecogniz-ing basic English sen-tences). Still, it is not quite clear how all these systems, which compute information auto-matically, integrate all this information to get to the level of a unified conscious cognitive state. Above all, it is not clear how exactly we should understand the relationship between physical (neural) and human cognitive phe-nomena in this perspective, especially when one considers some particular features of human cognitive phenomena.21
One way to deal with the problem of how these phenomena are related would be to en-dorse some kind of ontological neuro-cognitive reductionism and maintain that all aspects of our human mental life can be re-duced to physicochemical neural processes. As Gazzaniga acknowledges, many neuroscien-tists endorse an ontological reductionist
posi-tion of this kind. In their opinion, future ad-vances in understanding how the brain works will «reveal all one needs to know about how the brain enables the mind».22 However, even though Gazzaniga agrees with these basic tenets, i.e. that we are nothing but bio-logical machines living in a physical world,23 he does not consider it necessary to subscribe to an epistemological neuro-cognitive reduc-tionist account. Ultimately, he attempts to de-fend a kind of ontological monist physicalism, but at the same time an epistemological non-reductionist position, i.e. roughly, he believes that explanations of human cognitive phe-nomena have some autonomy regarding ex-planations for neural activity.
Gazzaniga’s view on epistemological neu-ro-cognitive reductionism is not entirely clear nor is it articulated in detail.24 However, he often describes this kind of reductionism using terms such as “derivation” or “predic-tion”.25 Given his most frequent and clear characterizations of epistemological reduc-tion and his counter-arguments against this reductionist position, the most plausible in-terpretation is that Gazzaniga implicitly re-lies on the traditional and influential theory of scientific reduction developed by Ernest Nagel.26 In this view, reductions are achieved by a deduction (or derivation) of a law or theory (T2) from a more fundamental law or theory (T1). The deduction provides a deduc-tive-nomological explanation, whereby T1 explains the phenomena explained by T2. In Nagel’s view, there are two kinds of theoreti-cal scientific reduction: between theories in the same domain of explanation and between theories in different domains. The first is called “homogeneous” and the second “het-erogeneous” reduction. When theories in play cover the same domains (are homoge-neous), the reduction can be achieved rela-tively easily because both theories are dealing with the same terms. However, when theo-ries in play cover what appears to be different domains (are heterogeneous – as in the case of neuroscience and psychology/cognitive science) and use different terms, then the
re-duction is more difficult to achieve. To real-ize a proper heterogeneous reduction follow-ing this theory, it is necessary to establish “bridge principles” that make the reductive links between the terms possible. Nagel does not specify the exact nature of these princi-ples, which could be characterized in terms of conditionals, bi-conditionals or identities. It is indeed in terms of identities (e.g. certain men-tal states and certain neural states are identi-cal, that is, they refer to the same thing)27 that Gazzaniga understands the reduction relation. As he maintains: «I do not think that brain-state theorists, those neural reductionists who hold that every mental state is identical to some as-yet-undiscovered neural state, will ever be able to demonstrate it».28
Gazzaniga identifies, however, three main difficulties involved in such epistemological reductions of cognitive to neural theories. The first difficulty is due to the argument from multiple realizability.29 This argument was initially formulated by Hilary Putnam30 and later accepted and developed by, among others, Fodor.31 It is a well-known argument that is most commonly put forward against the traditional psycho-neural type identity theory, but which also applies to Nagel’s tra-ditional intertheoretical reductionist ap-proach construed by means of using type identity relations as bridge laws. The argu-ment holds that the same types of cognitive phenomena can be realized by different types of physical neural systems. In this way, the human cognitive state of “being in pain”, for instance, is not necessarily identical to, let us say, a specific type of connection between spe-cific types of neurons in the human brain, be-cause it could be the case that this type of mental state could be realized in the brains of organisms that do not have this specific type of neural connection and yet the mental state of being in pain would remain the same (e.g. the brain of an octopus could realize the same pain the human brain does).32 Therefore, types of cognitive phenomena cannot be iden-tical to types of neural phenomena and cannot be ontologically reduced to them in this way.
Accordingly, in Gazzaniga’s view, systems in the brain are «flexible and largely inde-pendent».33 They can be changed or re-placed, with the sole requirement that they be compatible with the general architecture of the whole greater system of which they form a part. For example, if a region in this brain is damaged, it could be replaced by an artificial component, as long as the new sub-system retains the same functional role as the original component (having the same causal role, producing the same effect) and is com-patible with the entire structure of the brain. And this substitution could also be made if the artificial part were in some respects dif-ferent from the original subsystem, i.e. the original neurons and their respective connec-tions. In this way, a mental function related to a specific behavior can be implemented or realized in many different forms and thus – on the basis of the argument of the multiple realizability – they cannot be considered as reducible through identities to some specific neural processes. This ontological point is not incompatible with ontological monist physicalism (it just leads to token-token neu-ro-cognitive physical ontological relations), but it has important epistemological implica-tions for the relaimplica-tions between neural and cognitive theories. Given the multiple real-izability of human cognitive phenomena and the impossibility of constructing bridges be-tween neural and cognitive concepts, it is also not possible to achieve an intertheoretical reduction along these lines. Thus, the multi-ple realizability of human cognitive phenom-ena leads to epistemological human neuro-cognitive non-reductionism.
The second difficulty for the epistemolog-ical neuro-cognitive reductionist, in Gazza-niga’s view, is the fact that the brain is a dy-namical complex system whose effects can never be completely determined or predicted. This kind of system is defined as something composed by many different subsystems that interact. This interaction is so complex that it produces genuinely new properties and thereby states which are «greater than the
sum of their parts»34and cannot be reduced to the operations of the component parts. According to Gazzaniga, it is a «basic prin-ciple in experimental science» that «no measurement is infinitely precise», and there will «always [be] imprecision», therefore any measure will always include a «degree of uncertainty in the value».35 For Gazzaniga, such measurements are indeed impossible “in principle”: «Uncertainty is present because no matter what measuring device is used, it has a finite precision and, therefore, impreci-sion, which can never be eliminated com-pletely, even as a theoretical idea».36 And since the initial conditions of a system cannot be measured with complete accuracy, i.e. the initial measurement will always have a degree of uncertainty, the results derived from that measurement will also be uncertain.
If we now apply this reasoning to the rela-tion between neural and cognitive phenome-na, we need to conclude, following Gazzani-ga’s reasoning, that it is impossible to predict cognitive states and processes just by looking at physicochemical neural interactions and laws. Systems with great complexity in their organization sometimes interact in a com-pletely different form than simpler systems (e.g. with nonlinear causal interactions, such as positive and negative feedback loops). These complex interactions produce new properties that behave in accordance with laws that are radically different from those related to the components of the complex system and which cannot be deduced or de-rived from the laws associated with those components. This is the reason why complex systems are unpredictable, and therefore, ac-cording to Gazzaniga, irreducible.37
With this account of the nature of human cognitive phenomena, the author does not intend to give up on the monist ontological physicalism he endorses. Even if these new properties somehow emerge in whole human neuro-cognitive systems given the complex interactions between their parts, they are not something different from, or beyond, physi-cal nature. No mysterious new cognitive
causal powers are available for these new cognitive capacities, states or properties. However, this ontological account and its epistemological implications compromise the epistemological human neuro-cognitive re-duction, as characterized by Gazzaniga. For a precise derivation between laws in the two different domains (neural and cognitive) would not be possible. Such a derivation would not be possible because laws relating to the behavior of the whole system would not follow from the laws relating to the be-havior of the parts taken in isolation. Given that there is no derivation of neural and cog-nitive laws, there is no epistemological hu-man neuro-cognitive reduction either. For Gazzaniga, a similar problem is found in the domain of physics, since Newton’s laws of classic mechanics cannot be derived from laws related to quantum mechanics.38
Finally, Gazzaniga suggests that explana-tions that span multiple levels are needed in order to explain how neuro-cognitive process-es work.39 Natural phenomena are complex and they cannot be completely explained at a single level, e.g. at some fundamental physical level. Instead, they seem to be organized in many levels that go – to consider physical and biological descriptions only – from elementary particles to atoms and molecules, from mole-cules to cells, and from unicellular organisms to complex multicellular organisms with com-plex organs, like the brain. The brain itself can be described at different levels of organiza-tion, e.g. at the level of molecules, neurons, neural connections, neural networks, neural systems, etc. At these various levels, there are, for Gazzaniga, different types of modules or systems, as well as different types at the same level. Interactions occur both intra-level i.e. among systems at the same level (for example, among neurons in the brain) and inter-level, i.e. among systems and their components (subsystems) at different levels (for example, among atoms or molecules and the whole structure of a specific neuron).
According to Gazzaniga, reality itself is organized in multiple levels and this leads to
the irreducibility of scientific explanations in the various sciences: we will always have mul-tiple sciences which are largely autonomous from each other. In contrast with the inter-theoretical account of scientific reduction, he argues that theories at a higher-level of ex-planation cannot simply be deduced from theories at a lower-level of explanation (e.g. one cannot deduce theories based on biologi-cal laws from theories based on chemibiologi-cal laws). There are different levels of explana-tion, each with its particularities and rele-vance, and one level cannot be simply logical-ly derived from another as was hypothesized by advocates of intertheoretical epistemolog-ical reductionism. This applies not only to biology or chemistry, but also to cognitive neuroscience and cognitive science. Accord-ing to Gazzaniga, multi-level explanations as well as the unpredictability and multiple real-izability of human cognitive phenomena are incompatible with an intertheoretical human neuro-cognitive reductionist framework. █ Gazzaniga’s theory as a reductionist theory
In the previous section we considered why Gazzaniga takes epistemological neuro-cognitive non-reductionism to be a preferable account to epistemological neuro-cognitive re-ductionism. The former position, in his view, is consistent with: (1) the multiple realizability of human cognitive phenomena; (2) the unpre-dictability of cognitive phenomena produced by dynamical complex neural systems; (3) a multi-level account of scientific explanation.
The issue concerning his conclusion is, therefore, whether epistemological neuro-cognitive reduction is necessarily jeopardized by multiple realizability and the unpredicta-bility of cognitive phenomena, and also whether it is necessarily incompatible with any kind of multi-level account of cognitive phenomena. To address these points, we need to better clarify the controversial notion of reduction. That Gazzaniga choose to base his view on a critique of the intertheoretical Nagelian theory and psycho-neural type
identity theory is reasonable to the extent that these theories are still highly influential in present debates. However, many accounts of reduction have been proposed in philoso-phy of science and philosophiloso-phy of mind espe-cially since the late 1920s, and there are still today many intense debates about which ac-count best describes scientific practice and the possible reduction of cognitive phenomena.40
Recently, a new comprehensive theory for understanding scientific practice especially in the life sciences, including cognitive neuro-science, has been proposed. The application of this theory to cognitive neuroscience pro-vides the necessary material to counter Gaz-zaniga’s arguments.
This new comprehensive theory is part of what can be called neo-mechanist
philoso-phy,41 a new philosophical body of funda-mental theories that have been highly influ-ential in contemporary philosophy of science and are advocated by many scholars.42 A broadly neo-mechanist trend in philosophy of science gained more prominence approx-imately in the last quarter of the twentieth century. It can be divided into two major traditions.43 The first tradition is represented most prominently by the works of Wesley Salmon44 and Phil Dowe,45 and makes use of the concept of mechanism primarily to dis-cuss causation. The second tradition is repre-sented most prominently by the works of William Bechtel and Robert Richardson, as well as Stuart Glennan, Peter Machamer, Lindley Darden and Carl Craver.46 These au-thors, instead, use the notion of mechanism to describe how scientific explanations work, especially when they need to account for complex systems (e.g. in the biological sci-ences). In this paper, I focus solely on the second tradition, which is better suited to a discussion of how scientific explanations in cognitive science and neuroscience might best relate to each other.
This tradition, to the extent that it is con-cerned with the biological sciences, can be re-garded as providing two major theories about biological phenomena and scientific
explana-tions. The first theory is ontological and it clarifies what a complex system biological
mechanism (CSB-mechanism) is: it can thus be
called the mechanistic theory of biological
sys-tems (MTBS). The second theory is concerned
with epistemology and it clarifies how expla-nations of CSB-mechanisms should be con-structed: it can be called a mechanistic theory
of scientific explanations in the biological scienc-es (MTSEBS).
Given that neo-mechanists consider cogni-tive neuroscience to be a biological science, there are also two versions of this theory ap-plied to cognitive neuroscience. The first clari-fies what a complex system neuro-cognitive
mechanism (CSNC-mechanism) is: thus, it is a mechanistic theory of neuro-cognitive systems
(MTNCS). The second clarifies how CSNC-mechanisms should be explained: thus, it is a
mechanistic theory of scientific explanations in
cognitive neuroscience (MTSECN).47 The
clearest and most detailed accounts related to cognitive neuroscience are given by William Bechtel and Carl Craver.48
Bechtel and Craver make indeed a great ef-fort to show that the classical theory of scien-tific intertheoretical reduction proposed by Nagel is wrong because it relies on the
deduc-tive-nomological theory of scientific explanation
(DNTSE),49 i.e. on the idea that something can be explained only if it can be deduced from some universal laws of nature.50 Contra-rily, MTSEBS and thus MTSECN do not focus on logical deduction and universal laws, which are typically not useful for the concrete work of biologists and cognitive neuroscien-tists. In contrast, the starting point of these theories is the general claim that scientists working in the life sciences formulate explana-tions in terms of “mechanisms”.
Accordingly, a mechanism, roughly put, is «a structure performing a function in virtue of its component parts, component opera-tions, and their organization».51 The core idea is that mechanisms are hierarchical sys-tems that are made of components (working parts) and their operations; the working parts perform operations and interact
causal-ly with other parts of the mechanism. The general behavior of the whole system is a re-sult of the specific organization of the com-ponents and their interactions. A mechanistic explanation should describe how the orga-nized functioning of the mechanism is re-sponsible for the phenomenon to be ex-plained. According to MTNCS, neuro-cognitive complex systems, as investigated by cognitive neuroscience, must be understood as complex neural systems that process in-formation in a particular way. Bechtel points out that cognitive neuroscientists «have ap-pealed to systems-level understanding of the brain as providing the appropriate point of connection to the information processing ac-counts advanced in psychology».52 Thus, the theory endorses ontological monist physical-ism, as Gazzaniga does. Ultimately, all phe-nomena in reality have a physical nature, in-cluding (human) cognitive phenomena. As Craver and Tabery point out, many neo-mechanists accept the «central ideas that mo-tivate a broadly physicalist world-picture».53 MTSECN, on the other hand, provides the norms for the scientific explanations of such neuro-cognitive complex systems.
Bechtel confirms that «from the point of view of mental activity» this approach is re-ductionist, and he calls it “mechanistic
reduc-tion”.54 The author’s mechanistic approach
«emphasizes the need to identify all (or at least the major) operating parts of the mech-anism responsible for the phenomenon of in-terest and to understand the way they are or-ganized and how their operations are orches-trated to realize the phenomenon».55 Ac-cordingly, in a mechanistic neuro-cognitive epistemological reduction, properly con-structed, one must first of all describe the phenomenon to be explained and then local-ize the mechanism responsible for producing the phenomenon. Once this is done, the pro-cess of decomposition can start: this is aimed at understanding how the components work, how their operations are performed and how they are organized in order to produce the phenomenon. Since complex biological
mechanisms are composed of parts and their organization, the decomposition of a whole mechanism and its explanation can span multiple levels. A mechanistic epistemologi-cal neuro-cognitive reduction is achieved when a model of a set of mechanisms and their interactions at a lower-level can fully account for the causal processes of the whole mechanism at a higher-level, even when this whole mechanism is described with a com-pletely different vocabulary – as long as the terms can be properly related.
One of the best examples of a concrete ap-plication of MTNCS and MTSECN to specific cognitive phenomena concerns memory, which has been traditionally an object of study in the field of psychology. Functional analyses of the memory system reveal the existence of many memory subsystems such as, e.g., short-term memory, long-term memory, phonological memory, visuo-spatial memory, semantic memory, episodic memory, etc. as well as dif-ferent processes related to them, such as encod-ing, storage, consolidation, retention, and re-trieval.56 The identification of sub-functions of a system is the kind of work that cognitive psy-chologists do most of the time. In mechanistic explanations this is called functional decompo-sition. Due to the great complexity of the memory system it is more useful to seek expla-nations for each sub-function first, and then try to understand how they are related considering the mechanism of memory as a whole.
One of the best understood phenomena in the memory system is memory consolida-tion. Roughly put, this is the phenomenon of transforming short-term memories (which are labile and easy to disrupt) into long-term memories, which are robust and enduring and permit the organism to remember im-portant events for a longer period of time and modify its behavior accordingly.57 To explain this phenomenon, all the relevant re-gions in the brain responsible for the func-tions that compose the neuro-cognitive mechanism of memory consolidation, includ-ing all relevant mechanistic levels of decom-position, must be identified through the
pro-cess of localization, i.e. all the particular component parts and component operations of the whole mechanism must be determined. Finally, the causal processes and causal inter-actions within the mechanism’s functions need to also be understood, i.e. the general organization of the mechanism.
The explanation starts at the highest level of the whole mechanism. At this level, it is necessary to correctly identify the entire neu-ral network that is responsible for memory consolidation (recent studies show that this includes the hippocampus and other particu-lar areas in the brain).58 Secondly, it must be established whether this large neural system is indeed all that is relevant for the explanation of the phenomenon. For this, techniques such as fMRI can be useful, since they can specify what regions in the brain are activated during the performance of a specific psychological task. However, a mechanistic explanation at this level also needs to clarify how the neural network encodes and decodes new memory episodes through information processing and computational operations and how these pro-cesses produce and affect, for instance, the dif-ferent degrees of consolidation that character-ize the memories under investigation.
Once this has been clarified, a second lev-el of explanation must be provided in which the large neural system is decomposed into particular sub-neural systems localized in more specific regions. Here the goal is to un-derstand the information processing and computational operations (e.g. spiking pat-terns in populations of neurons) of these smaller neural networks and how they con-tribute to the performance of the whole mechanism composed by such neural-nets.
Moreover, a further stage of decomposi-tion must be reached: the processes underly-ing memory at an inter-cellular level. The ex-planation at this particular level aims at de-scribing the components of a particular neu-ral network and at understanding how small numbers of neurons operate (for example, how they depolarize and fire in the process of propagating action potentials, or how they
are responsible for synaptic processes, neuro-transmitters being released, and so on). Here it is possible to measure the spiking rates of neurons, or spiking frequency and to record neural activity in general. There is already sufficient evidence to clarify, for example, the process of long-term potentiation (LTP). The process is understood as: «the persisting enhancement in the response of a postsynap-tic cell to an input from a presynappostsynap-tic cell when the postsynaptic cell has readily spiked after inputs from the presynaptic cell».59 Many authors consider this process to be central for the consolidation of memories at a purely physiological level.
Finally, the explanation can go to another lower mechanistic level: the intra-cellular and molecular level. At this level, the description is in terms of the activity of the relevant pro-teins, molecules and ions, such as, for in-stance: N-methyl-D-aspartate (NMDA), pro-tein kinase A (PKA), cyclic adenosine mono-phosphate (cAMP), α-amino-3-hydroxy-5-methyl-4-isoxazole proprionic acid (AMPA), Ca2+, Na+ and K+ ions.60 These proteins and molecules are considered to be im-portant for the processes of LTP, i.e. compo-nents in the mechanism of LTP, which is a sub-sub-subsystem in the whole mechanism of memory consolidation and plays a role that, together with the operations happening at higher levels, allows for this phenomenon to take place. As one can see, this kind of ex-planation «exhibits a progression from be-havioral-level characterization of memory consolidation to identification of important components in the process at progressively lower levels».61 In the end, all levels are equally important in achieving a complete multilevel mechanistic explanation of the particular phenomenon.
According to Bickle, some scientific ex-periments already present evidence for estab-lishing the connection between a molecular and cellular mechanism and a particular be-havior that indicates a cognitive function, as in certain cases of memory consolidation.62 In one scientific experiment, a mouse in
which the cellular transcription factor CREB was “knocked-out” had intact short-term memory on many rodent memory tasks, but showed a great decrease in memory capacity in comparison with the control for the long-term memory versions of these tasks. In an-other experiment CREB was increased in a small population of neurons in a manipulated mouse. This led to an increase of memory consolidation measured by fear conditioning behavior in the modified mice. Since CREB has traditionally been considered to be impli-cated in the induction of LTP, ultimately it is arguably the presence of CREB that is doing the most central bottom-level causal work. In fact, CREB is part of a particular molecular mechanism that involves cAMP, PKA and CREB, which leads to LTP. Bickle claims that blocking any step of this mechanistic process «virtually eradicates memory consolidation, while enhancing steps can lead to faster and stronger consolidation».63 Thus, it is these smaller molecular mechanisms that, taken together, explain the behavior of the whole mechanism of memory consolidation in these particular cases. This kind of explanation supports the integration of neural hierar-chical compositional mechanisms at different levels. This is the reason why it is not only compatible with the scientific explanations found in the field of cognitive neuroscience but it is also the most appropriate approach for the explanatory and integrative goals of the field: i.e. to integrate neural and cognitive scientific theories and to describe (human) cognitive processes through the activities of neural systems. However, this is also a neuro-cognitive reductionist explanation, since the ‘lower-level’ mechanisms and their interac-tions at this lower-level can always account for the causal interactions of “higher-level” mech-anisms. Lower-level neuro-cognitive mecha-nistic models can completely account for the causal explanation described (in different terms) by neuro-cognitive mechanistic models at a higher-level. Thus, there is a mechanistic epistemological neuro-cognitive reduction.
This neuro-cognitive mechanistic
onto-logical and epistemoonto-logical reductive frame-work is able to respond to all the three ex-planatory requirements that Gazzaniga points out. Firstly, mechanistic epistemologi-cal neuro-cognitive reduction provides an answer for the multiple realizability of cogni-tive phenomena, as framed by Gazzaniga.64 More particularly, the mechanistic theory applied to cognitive neuroscience avoids the general problem of multiple realizability of cognitive phenomena because it neutralizes the argument, as provided by the author. In Bechtel’s view, evidence from neuroscience (comparative studies with brain-damaged animals, PET and fMRI) shows that there are many relevant similarities in brain areas across species; he argues that if mental states were described in terms as fine-grained as those used to describe neural states (e.g. if we hypothesize that a specific neural mechanism produces a specific state of pain, rather than pain in general), the multiple realizability of cognitive phenomena becomes less plausi-ble.65 A single neuro-cognitive mechanism responsible for producing a specific cognitive state (e.g. a human pain state, or a dog pain state), with all its particularities, can be local-ized; hence, local reductions through (heuris-tic) identity relations (involving more equivalently grained states) are plausible. In other words, as Bechtel points out, if one uses the same standards of typing for cognitive and neural phenomena, in terms of fine and coarse grain, then «types might range across species and enable scientists to claim that the same type of mechanism in different species produces the same type of [cognitive] phe-nomena».66 In this sense, it may be possible to establish reductive identity relations: «Type identity claims are core to the prac-tice of mechanistic explanation in biology [and cognitive neuroscience] and are not jeopardized by the philosophical claims of multiple realization».67 Since cognitive func-tions and processes can thus be identified with functions and processes of neural mech-anisms, there can be epistemological neuro-cognitive reduction between neuro-cognitive
theo-ries and models of cognitive systems and neuroscientific theories and models of neural mechanisms. This is an argument in contrast with Gazzaniga’s view that neuro-cognitive type identities cannot be established at all. Since in some particular cases, for some par-ticular neuro-cognitive processes, identities are arguably plausible, as in some particular cases of memory consolidation,68 Gazzaniga’s point is misleading. Thus, Bechtel’s approach remains more plausible than Gazzaniga’s, even if the general issue of multiple realiza-bility is still quite controversial.
The second argument raised by Gazzani-ga aGazzani-gainst neuro-cognitive epistemological reductionism is concerned with the idea that this kind of reduction necessarily implies predictability.69 In his line of reasoning, if one has all the relevant knowledge of the el-ements in the deduction of a certain behav-ior, s/he must be able to anticipate (predict) all the aspects of this behavior; in the same way that Newton’s classical laws of mechan-ics and his universal law of gravitation or Einstein’s theory of general relativity can, with a high level of precision, predict the mo-tion of the planets. However, since the theory of mechanistic neuro-cognitive epistemolog-ical reduction does not require logepistemolog-ical deduc-tion from fundamental to less fundamental laws and their related properties, nor the kind of high level predictability that results from this, it is compatible with a high degree of indetermination. In other words, with re-spect to mechanistic theory it is irrelevant if at a certain time an event or a set of events occur that cannot be deduced or predicted with great accuracy by these related theories on the basis of the initial state of the system due to the inaccuracy of measures taken at the start or given the complexity of the rela-tions between its parts. What really matters for a mechanistic epistemological neuro-cognitive reduction is whether it is possible to describe these properties in causal, func-tional, computafunc-tional, neural or any other mechanistic physical language – and then to identify these mechanisms and their
compo-nents as far as possible (i.e. at least the most relevant ones that contribute causally to a given effect), ultimately producing a scien-tific theory to account for them at a lower-level. This does not mean that the mechanistic theory gives up on predictability once and for all. It simply accounts for the fact that the more dynamical and complex a system is, the less its outcomes will be predictable. As the examples of particular kinds of memory con-solidation given by Bickle show,70 descriptions of the behavior produced by the lower-level molecular mechanisms are responsible for the explanation of the causal effects of the whole higher-level mechanism, even though there is no deduction or laws involved.
Concerning the last argument developed by Gazzaniga, it is not even clear that Nagel’s theory cannot account for multiple levels of scientific explanation. Nagel intended to de-velop a theory to integrate scientific activity and, thus, relate all the domains of science having physics as its basis. However, Nagel had a very particular notion of reduction in mind. He was concerned with theories in a given scientific domain and across scientific domains. Theories in his view were construct-ed as a set of laws and statements about reali-ty. Nagel does not assert in his major work that classic physics should literally reductively explain biological phenomena or psychologi-cal phenomena. Even though he argued for scientific reduction across all of the sciences, his view appears to still be compatible with an account of multiple levels of natural phenom-ena and scientific domains, as long as these levels are properly related. But leaving aside this issue with respect to Nagel, the mechanis-tic theory applied to cognitive neuroscience is built on the idea that we need to rely on dif-ferent hierarchical “levels of explanation”.71 According to Machamer, Darden and Craver, the main interest of neo-mechanists is a multi-level explanation since the life sciences have
per se a multilevel character. «It is the
integra-tion of different levels into productive rela-tions that renders the phenomenon intelligible and thereby explains it».72 However, when
lower-levels can account for the causal pro-cesses of higher-levels in a given neuro-cognitive complex mechanism, mechanistic scientific explanations in cognitive neurosci-ence become epistemologically reductive, even though higher-level explanations are still ac-cepted as adequate descriptions of the phe-nomena in a different language.
Therefore, Gazzaniga’s view is entirely compatible with a particular form of neuro-cognitive ontological and epistemological re-ductionism, namely neo-mechanistic neuro-cognitive reductionism. Gazzaniga’s view concerning neuro-cognitive phenomena and related explanations is thus misleading. Be-cause of this, MTNCS and MTSECN pro-vide a more consistent and coherent view for the contemporary field of cognitive neurosci-ence, given that these theories are compatible with the explanatory strategies and ultimate explanatory aims of this scientific domain. █ Concluding remarks
Gazzaniga deserves credit for many ad-vances in our understanding of how particular brain mechanisms operate; more importantly, he is one of the most significant pioneers in the field of cognitive neuroscience, and a lead-ing scientist interested in research about the neuro-cognitive relation. In his attempt to construct a more systematic theory about the relationship between human brain and cogni-tion based on his achievements and more gen-erally on some of the achievements of cogni-tive neuroscience, he comes to the broad con-clusion that any epistemological reductionist account concerning explanations of human cognition must be rejected. In his view, epis-temological neuro-cognitive reductionism is incompatible with the multiple realizability and unpredictability of human cognitive phe-nomena and also with the idea that explana-tions in neuro-cognitive sciences are intrinsi-cally multi-level.
However, to argue for epistemological neuro-cognitive non-reductionism on these general grounds turns out to be misleading.
As I show in this paper, recent neo-mechanistic theories are able to account for these arguments, as framed by Gazzaniga, and still offer a robust form of ontological and epistemological neuro-cognitive reduc-tion. Thus, this form of reduction is perfectly in line with Gazzaniga’s view on human cog-nition. More importantly, this theory has the additional advantage of providing the most accurate description so far of the kind of sci-entific explanations constructed in contem-porary cognitive neuroscience, thereby offer-ing foundational support for its explanatory aims. Therefore, Gazzaniga’s claim that un-derstanding the relation between neurologi-cal and cognitive processes in human cogni-tion means relying on a non-reductive expla-nation is untenable. On the contrary, the ap-proach that appears to be most compatible with the explanatory aims and strategies of contemporary cognitive neuroscience is the neo-mechanist ontological and epistemologi-cal neuro-cognitive reductionist account.
It is important to emphasize that this analysis does not intend to provide a defense of MTCNS and MTSECN. Indeed, these theories have many limitations that could not be addressed in this paper.73 The point of the present work is solely to argue that, given the general ontological monist physicalist fra-mework that Gazzaniga accepts and the ar-guments against epistemological (human) neuro-cognitive reductionism that he pro-vides, his positions on the epistemological irreducibility of human cognitive phenomena do not hold up. Accordingly, this paper at-tempts to offer a genuine contribution to de-bates concerning the plausibility of funda-mental theories of cognition and neuro-cognitive integration in neuro-cognitive neurosci-ence.
█ Notes
1 For instance, Gazzaniga established the Cognitive Neuroscience Institute, founded the Journal of Cog-nitive Neuroscience and cofounded the CogCog-nitive Neuroscience Society. He is also first author of the
popular textbook, Cognitive Neuroscience: the Biol-ogy of the Mind.
2 See P.A. REUTER-LORENZ, K. BAYNES, G.R. MANGUN,E.A.PHELPS (eds.), The Cognitive
Neuro-science of Mind: A Tribute to Michael S. Gazzaniga, MIT Press, Cambridge: (MA) 2010.
3 See M.S.GAZZANIGA, The Bisected Brain, Apple-ton Century Crofts, New York 1970; M.S.G AZZA-NIGA, Tales from Both Sides of the Brain: A Life in Neuroscience, Ecco, New York 2015; M.S.G AZZA-NIGA,J.E.LEDOUX, The Integrated Mind, Plenum Press, New York 1978.
4 See M.S. GAZZANIGA, The Ethical Brain, Dana Press, New York 2005.
5 See M.S.GAZZANIGA, Who’s in Charge? Free Will and the Science of the Brain, Ecco, New York 2012- II edition.
6 See M.S.GAZZANIGA, Mind Matters: How Mind and Brain Interact to Create our Conscious Lives, Houghton Mifflin, Boston 1988; M.S.GAZZANIGA, Nature’s Mind: The Biological Roots of Thinking, Emotions, Sexuality, Language and Intelligence, Basic Books, New York 1992; M.S. G AZZANI-GA, The Mind’s Past, University of California Press, Berkeley (CA) 2000; M.S. GAZZANIGA, Human: The Science Behind What Makes Us Unique, Ecco, New York 2008.
7 See M.S.GAZZANIGA, Social Brain: Discovering the Networks of the Mind, Basic Books, New York 1985. 8 I use here “cognition” and “cognitive” as synony-mous of “mind” and “mental”, as it is common in the specialized literature. Moreover, I am interest-ed here particularly in the human adult normally functioning cognition.
9 See B.J.BAARS,N.M.GAGE, Fundamentals of Cog-nitive Neuroscience, Elsevier, Oxford (UK) 2013; M.S.GAZZANIGA,R.B.IVRY,G.R.MANGUN, Cogni-tive Neuroscience: The Biology of the Mind, Norton, New York 2014 - IV edition; K.N.OCHSNER,S.M. KOSSLYN (eds.), The Oxford Handbook of Cognitive
Neuroscience. Volume I, Oxford University Press, Oxford (UK) 2014; J.WARD, The Student’s Guide
to Cognitive Neuroscience, Psychology Press, New York 2015 – III edition.
10B.J.BAARS,N.M.GAGE, Fundamentals of Cogni-tive Neuroscience, cit., p. 3.
11J.WARD, The Student’s Guide to Cognitive Neuro-science, cit., p. xi, see also pp. 1-2.
12 See J.KIM, Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation, MIT Press, Cambridge (MA) 1998; J.KIM, Physi-calism, or Something Near Enough, Princeton
Uni-versity Press, Princeton 2005; D.PAPINEAU, The Rise of Physicalism, in: C. GILLETT, B. LOEWER
(eds.), Physicalism and Its Discontents, Cambridge University Press, Cambridge 2001, pp. 3-36; D. PAPINEAU, Naturalism, in: E.N.ZALTA (ed.), The Stanford Encyclopedia of Philosophy (Winter 2016 Edition), URL: https://plato.stanford.edu/ ar-chives/win2016/entries/naturalism/; D. STOLJAR, Physicalism, Routledge, London/NewYork 2010; D. STOLJAR, Physicalism, in: E.N. ZALTA (ed.), The
Stanford Encyclopedia of Philosophy (Spring 2015 Edition), URL: http://plato.stanford.edu/archives/ spr2015/entries/physicalism/.
13 See e.g. J.BICKLE, Philosophy and Neuroscience: A Ruthlessly Reductive Account, Kluwer Academic Press, Norwell (MA) 2003; J. BICKLE, Reducing
Mind to Molecular Pathways: Explicating the Re-ductionism Implicit in Current Cellular and Molecu-lar Neuroscience, in: «Synthese», vol. CLI, n. 3, 2006, pp. 411-434; J. BICKLE, A Brief History of Neuroscience’s Actual Influences on Mind-Brain Re-ductionism, in: S.GOZZANO,C.S.HILL (eds.), New
Perspectives on Type Identity: The Mental and the Physical, Cambridge University Press, Cambridge (UK) 2012, pp. 88-110; P.S. CHURCHLAND, Can Neurobiology Teach Us Anything about Conscious-ness?, in: «Proceedings and Addresses of the American Philosophical Association», vol. LXVII, n. 4, 1994, pp. 23-40; F.CRICK, The Astonishing Hypothesis, Scribner’s, New York 1994; J.G.K EM-MENY,P. OPPENHEIM, On Reduction, in: «Philo-sophical Studies», vol. VII, n. 1-2, 1956, pp. 6-19; E.NAGEL, The Structure of Science. Problems in the Logic of Explanation, Harcourt, Brace & World, New York 1961; P.OPPENHEIM,H.PUTNAM, The Unity of Science as a Working Hypothesis, in: G. MAXWELL,H.FEIGL,M.SCRIVEN (eds.), Concepts, Theories, and the Mind-Body Problem, Minnesota University Press, Minneapolis 1958, pp. 3-36. 14 See K.AIZAWA, C. GILLETT, The Autonomy of Psychology in the Age of Neuroscience, in: P.M. IL-LARI, F. RUSSO, J. WILLIAMSON (eds.), Causality in the Sciences, Oxford University Press, New York 2011, pp. 202-223; J.A.FODOR, Special Sciences, in:
«Synthese», vol. XXVIII, n. 2, 1974, pp. 97-115; J.A.FODOR, Special Sciences: Still Autonomous After All These Years, in: «Philosophical Perspectives», vol. XI, 1997, pp. 149-163; H.PUTNAM, The
Na-ture of Mental States (1967), in: H.PUTNAM, Mind Language and Reality: Philosophical Papers, Vol-ume II, Cambridge University Press, Cambridge (UK) 1975, pp. 429-440; D.WEISKOPF, The
Func-tional Unity of Special Science Kinds, in: «British Journal for Philosophy of Science», vol. LXII, n. 2, 2011, pp. 233-258.
15 See M.S. GAZZANIGA, Who’s in Charge?, cit., p. 33. See also M.S.GAZZANIGA, Social Brain, cit., p. x; M.S.GAZZANIGA, Tales from Both Sides of the Brain, cit., p. 338 and 340.
16 See M.S.GAZZANIGA, Who’s in Charge?, cit., pp. 40-41; M.S.GAZZANIGA, Social Brain, cit., p. 18; M.S. GAZZANIGA, Tales from Both Sides of the
Brain, cit., p. 337.
17 For the distinction between phenomenal and ac-cess consciousness see N.BLOCK, On a Confusion about the Function of Consciousness, in: «Behavior-al and Brain Sciences», vol. XVIII, n. 2, 1995, pp. 227-247.
18 M.S.GAZZANIGA, Who’s in Charge?, cit., p. 66. 19 Ivi, pp. 84-85. See also M.S.GAZZANIGA, Tales from Both Sides of the Brain, cit., p. 150ff.
20 M.S.GAZZANIGA, The Ethical Brain, cit., p. xviii. 21 See F.ADAMS,K.AIZAWA, The Bounds of Cogni-tion, Wiley-Blackwell, Malden (MA) 2010; D. CHALMERS, Facing up to the Problem of Conscious-ness, in: «Journal of Consciousness Studies», vol. II, n. 3, 1995, pp. 200-219; J.A.FODOR, LOT 2: The Language of Thought Revisited, Oxford University Press, Oxford 2008; F. JACKSON, Epiphenomenal Qualia, in: «Philosophical Quarterly», vol. XXXII, n. 127, 1982, pp. 127-136; F. JACKSON, What Mary Didn’t Know, in: «Journal of Philoso-phy», vol. LXXXIII, n. 5, 1986, pp. 291-295; J. LEVINE, Materialism and Qualia: The Explanatory Gap, in: «Pacific Philosophical Quarterly», vol. LXIV, n. 4, 1983, pp. 354-361; T.NAGEL, What is it
Like to be a Bat?, in: «Philosophical Review», vol. LXXXIII, n. 4, 1974, pp. 435-456.
22 M.S.GAZZANIGA, Who’s in Charge?, cit., p. 4. 23 See ivi, p. 7
24 Sometimes he frames reduction in general as de-fending the thesis that «a complex system is noth-ing but the sum of its parts» (e.g. M.S.G AZZANI-GA, Tales from Both Sides of the Brain, cit., p. 335).
Probably Gazzaniga means by this that epistemo-logical reduction is the explanation of a property, function or behavior of a given whole system by means of the mere “sum” of the properties, func-tions or behaviors of its component parts. For in-stance, the total weight of a given physical object (e.g. a motorbike) is equal to the sum of the weights of its parts. Weight, however, is a property traditionally investigated in physics, it is a variable that can be completely quantified according to
standard methods of measurement, and “sum” here is simply mathematical addition. When we are talking about complex biological mechanisms, such as the brain, this reasoning about reduction is problematic, since, for instance, the mathematical addition of different physiochemical processes in neurons would never explain the processes that a neural network is capable of performing. The in-terrelations between the most interesting and rele-vant processes and functions here are not those of simple mathematical addition, as in the case of summing weights. They are relations of causation between physical activities. Indeed, no author has ever advanced any detailed account of epistemo-logical reduction in the literature along these lines. 25 See, e.g., M.S.GAZZANIGA, Who’s in Charge?, cit., pp. 123, 124,130, and 134. See also M.S.G AZZANI-GA, Tales from Both Sides of the Brain, cit., p. 350. 26 See E.NAGEL, The Structure of Science, cit., chap. 11. Evidently, this is not to say, at all, that Gazzaniga is a Nagelian. The reductionist account described by Gazzaniga is the account of a theory he wants to criticize. He constructs it using elements taken from Nagel’s theory, even if he is not aware of this. 27 The idea of establishing identity relations be-tween mental and neural states is actually a tradi-tional well-known ontological reductive program in philosophy of mind developed at the end of the 1950s, namely the classic psycho-neural type-type identity theory, whose advocates are mainly Place (see U.T.PLACE, Is Consciousness a Brain Process?, in: «British Journal of Psychology», vol. XLVII, n. 1, 1956, pp. 44-50), Feigl (see H.FEIGL, The “Men-tal” and the “Physical”, in: H.FEIGL,M.SCRIVEN,
G. MAXWELL (eds.), Concepts, Theories and the Mind-Body Problem, Minnesota University Press, Minneapolis 1958, pp. 370-497) and Smart (see J. SMART, Sensations and Brain Processes, in: «The Philosophical Review», vol. LVIII, n. 2, 1959, pp. 141-156). The central claim which these authors agree on is that some types (or kinds, or classes) of mental states or processes are (contingently) iden-tical with some types (or kinds, or classes) of neu-rophysiological states or processes.
28 M.S.GAZZANIGA, Who’s in Charge?, cit., p. 130. This interpretation of reduction is not the one that Nagel himself endorsed (see, e.g., R.C. RICHARD-SON, Functionalism and Reductionism, in: «Philos-ophy of Science», vol. XLVI, n. 4, 1979, pp. 533-558; K.SCHAFFNER, Ernest Nagel and Reduction, in: «Journal of Philosophy», vol. CIX, n. 8, 2012, pp. 534-565); still, this is the one that best describes
how Gazzaniga conceives of epistemological re-duction and that best explains why he considers this kind of reductionism to be an inacceptable stance for cognitive phenomena.
29 See, ivi, p. 131.
30 See H. PUTNAM, The Nature of Mental States (1967), in: H.PUTNAM, Mind Language and Reali-ty: Philosophical Papers, Volume 2, Cambridge Uni-versity Press, Cambridge 1975, pp. 429-440. 31 See J.A.FODOR, Special Sciences, in: «Synthese», vol. XXVIII, n. 2, 1974, pp. 97-115.
32 H.PUTNAM, The Nature of Mental States, cit., p. 436.
33 M.S.GAZZANIGA, Tales from Both Sides of the Brain, cit., p. 349.
34 M.S.GAZZANIGA, Who’s in Charge?, cit., p. 71. 35 Ivi, p. 109.
36 Ibidem.
37 Gazzaniga also conflates issues about causal sci-entific determinism, indeterminism and free will in this discussion (see M.S. GAZZANIGA, Who’s in Charge?, cit., chap. 4). However, the clearest and most important part of his argument concerning epistemological neuro-cognitive non-reductionism relies on the unpredictability of cognitive phenom-ena given the complex interactions of neural sys-tems. Thus, his discussions about causal determin-ism do not concern us here.
38 See M.S.GAZZANIGA, Who’s in Charge?, cit., p. 125: «The balls in my living room are made up of atoms that behave as described by quantum me-chanics, and when those microscopic atoms come together to form macroscopic balls, a new behavior emerges and that behavior is what Newton observed and described. […] you can’t predict [derive] New-ton’s laws from observing the behavior of atoms, nor the behavior of atoms from Newton’s laws». 39 M.S.GAZZANIGA, Tales from Both Sides of the Brain, cit., p. 344. See also M.S.GAZZANIGA, Who’s in Charge?, cit., p. 125.
40 See, e.g., A. BECKERMANN, H. FLOHR, J. KIM (eds.), Emergence or Reduction? Essays on the Pro-spects of Nonreductive Physicalism, De Gruyter, Ber-lin 1992; R.C.RICHARDSON, S.STEPHAN,
Reduc-tionism (AntireducReduc-tionism, Reductive Explanation), in: M. BINDER, N. HIROKAWA, U. WINDHORST
(eds.), Encyclopedia of Neuroscience, Heidelberg, Springer 2009, pp. 3395-3398; S.SARKAR, Models of
Reduction and Categories of Reductionism, in: «Syn-these», vol. XCI, n. 3, 1992, pp. 167-194; M. SIL-BERSTEIN, Reduction, Emergence, and Explanation, in: M. SILBERSTEIN, P. MACHAMER (eds.), The
Blackwell Guide to the Philosophy of Science, Black-well, Oxford 2002, pp. 80-107; R. VAN GULICK, Reduction, Emergence and Other Recent Options on the Mind/Body Problem: A Philosophic Overview, in: «Journal of Consciousness Studies», vol. VIII, n. 9-10, 2001, pp. 1-34; R.VAN RIEL,R.VAN G U-LICK, Scientific Reduction, in: E.N.ZALTA (ed.), The Stanford Encyclopedia of Philosophy (Fall 2015 Edi-tion), URL: http://plato.stanford.edu/archives/ fall2015/entries/scientific-reduction. Some of the most frequently debated views are, for example, reduction as translation (see R.CARNAP, The Unity of Science, Kegan Paul, Trench, Trubner, and Co, London 1934; O.NEURATH, Philosophical Papers, 1913-1946, Riedel, Dordrecht 1983); reduction as intertheoretical deduction/derivation (see E. NAGEL, The Structure of Science, cit.); reduction as intertheoretical explanation (see J.G.KEMMENY,P. OPPENHEIM, On Reduction, in: «Philosophical Studies», vol. VII, n. 1-2, 1956, pp. 6-19); Schaffner’s theory of reduction (see K.SCHAFFNER, Approaches to Reduction, in: «Philosophy of Sci-ence», vol. XXXIV, n. 2, 1967, pp. 137-147); new wave reduction (J. Bickle, Psychoneural Reduction: The New Wave, MIT Press, Cambridge (MA) 1998; C.HOOKER, Towards a General Theory of Reduction. Part I: Historical and Scientific Setting, in: «Dialogue», vol. XX, n. 1, 1981, pp. 38-59; C. HOOKER, Towards a General Theory of Reduction. Part II: Identity in Reduction, in: «Dialogue», vol. XX, n. 2, 1981, pp. 201-236; C.HOOKER, Towards a General Theory of Reduction. Part III: Cross-Categorial Reduction, in: «Dialogue», vol. XX, n. 3, 1981, pp. 496-529); ruthless molecular and cellu-lar neuroscientific reduction (J.BICKLE, Philosophy and Neuroscience, cit.; J.BICKLE, Reducing Mind to
Molecular Pathways, cit.; J.BICKLE, Real Reduction in Real Neuroscience: Metascience, Not Philosophy of Science (and Certainly Not Metaphysics!), in: J. HOHWY,J.KALLESTRUP (eds.), Being Reduced: New
Essays on Reduction, Explanation, and Causation, Oxford University Press, Oxford 2008, pp. 34-51; J. BICKLE, A Brief History of Neuroscience’s Actual In-fluences on Mind-Brain Reductionism, cit.; A.SILVA,
J.BICKLE, The Science of Research and the Search for Molecular Mechanisms of Cognitive Functions, in: J. Bickle (ed.), The Oxford Handbook of Philosophy and Neuroscience, Oxford University Press, Ox-ford/New York 2009, pp. 91-126); reductive logi-cal behaviorism (see G.RYLE, The Concept of Mind,
Hutchinson, London 1949); reduction as psycho-neural type identity relation (see U.T. PLACE, Is
